dotfiles

Safely review, sync, and locally update forked tmux plugins. Use whenever the user mentions tmux-upstream, tmux plugin forks, `prefix + U`, a `connorads/<plugin>` fork being commits behind upstream, asks whether a tmux plugin update is dodgy/compromised/safe, or asks to sync/update a forked tmux plugin. Default to review-only and ask before syncing unless the user explicitly requested automatic safe sync.

Catalogue of preferred linter rules, TypeScript flags, clippy thresholds, and architectural boundary checks for making bug classes and design drift mechanically impossible. Use when setting up linting in a new project, hardening an existing project, responding to a class of bug by encoding a rule, or deciding which linter to reach for on a given stack. Pairs with the `hk` skill which handles wiring hooks.

Design and write effective tests for behavioural changes, bug fixes, and refactors. Use when choosing a test layer, practising TDD, picking doubles/fakes, reducing brittle or flaky tests, refactoring safely, or applying property-based, snapshot/approval, differential/metamorphic, or contract testing. For coverage, thresholds, mutation testing, fuzzing, and CI/hook enforcement, use the test-coverage skill.

Create, update, validate, and submit Homebrew formulae (homebrew-core, built from source). Use when the user mentions a Homebrew formula, Homebrew/homebrew-core, adding/updating a formula, brew create, building from source, a build system in a brew context (cargo/rust, go, cmake, meson, autotools/configure, make, python virtualenv, node/npm, ruby gem), resource blocks, depends_on/keg_only/uses_from_macos, the mandatory test do block, bottles, livecheck, brew bump-formula-pr, or when asked to run brew audit --new / brew test / brew style for a formula. For macOS GUI apps and prebuilt binaries use the homebrew-cask-authoring skill instead.

Systematically audit, improve, and enforce test coverage, and gate test quality in CI — across any ecosystem (TypeScript, Python, Go, Rust). Use to raise coverage, set thresholds, audit gaps, manage exclusions, merge reports, wire coverage into CI/hooks, or add mutation testing and fuzzing as quality gates. Composes with the hk skill for pre-commit enforcement. For how to design and write good tests — property-based, snapshot/approval, differential, contract, flaky-test handling — use the testing skill.

Safely refresh the vendored third-party agent skills in this dotfiles repo. Use whenever the user wants to update, refresh, upgrade, or re-pull vendored skills (`skills update`), or asks to check whether a skill refresh is safe / dodgy / compromised before committing. `skills update` is an unauthenticated git clone with no quarantine, no signature, and no scan — and skill files are instructions injected into every agent session — so this skill gates each refresh by reading the diff and only auto-commits trusted-source, clean-diff updates.

Reference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST

Best practices for Remotion - Video creation in React

Download images from private GitHub issues and PRs using authenticated signed URLs.

Design, review, or refactor logging for a service so it works as an observability primitive — the canonical wide-event / structured-log pattern. Use when the user is setting up logging for a new service, adding or auditing middleware, building a logger config (structlog, pino, winston), enabling Cloudflare Workers observability or Analytics Engine, instrumenting OpenTelemetry logs, planning sampling strategy, proposing logging schema changes, or saying things like "our logs are noisy", "we can't find anything in the logs", "improve our logging", or "add proper logging". Do NOT trigger for single casual `console.log`/`print`/`logger.info` additions during unrelated work.

Scaffold a SUPERVISOR.md runbook for watching a long-running agent loop in tmux. Use when asked to "supervise a loop", "watch a loop", "babysit a loop", "set up a supervisor", or when a task-loop / rl run needs someone operating the harness around it.

Analyze OpenCode conversation history to identify themes and patterns in user messages. Use when asked to analyze conversations, find themes, review how a user steers agents, or extract insights from session history.

Create Product Requirements Documents (PRDs) that define the end state of a feature through iterative design interview. Use when planning new features, migrations, or refactors. Generates structured PRDs with acceptance criteria, testing strategy, and architectural decisions.

Channel the mental models, decision frameworks, and communication style of real experts (Steve Jobs, DHH, Rich Hickey, etc.) to approach problems the way they would. Use when the user says "summon", "channel", "what would [name] think", "ask [name]", or wants an expert perspective on a problem.

Scaffold a loop directory for automated agent task execution. Use when asked to "create a task loop", "set up a loop", "scaffold a loop directory", "prepare tasks for rl", or "set up automated execution" for a backlog. Takes an existing backlog and generates PROMPT.md (loop contract), run-log.md (execution history), and .gitignore for ephemeral loop-state.md.

Decompose input into a structured task backlog for automated agent loops. Use when asked to "create a task plan", "break this into tasks", "decompose this PRD", "turn this into a backlog", or "plan tasks from" any input source (PRD, Figma feedback, GitHub issues, user requirements). Also use when the user provides a PRD, design doc, or requirements and wants executable tasks, even if they don't mention "task plan" explicitly.

Control interactive CLIs (python, gdb, etc.) via tmux sessions - send keystrokes and scrape output

Audit, implement, and fix web accessibility with a screen-reader-first lens. Use when building or reviewing UI components, forms, dialogs, navigation, dynamic content, or any interactive element. Covers WCAG 2.2 AA compliance, ARIA patterns, keyboard navigation, focus management, and assistive technology compatibility (NVDA, JAWS, VoiceOver). Trigger on: "accessible", "a11y", "screen reader", "WCAG", "ARIA", or when adding any interactive UI.

Design or refactor software around clear boundaries, typed domain models, and testable flows. Use for substantial design work, hard-to-test code, domain modelling, ports/adapters, functional core / imperative shell, explicit error handling, observability, or when code structure is fighting the change.

Design holistic user experiences using systems thinking, service design, and psychological principles. Triggers on: UX design, user experience, journey map, service blueprint, user flow, wireframe, accessibility, WCAG, design critique, heuristic review, cognitive load, design thinking, holistic design, JTBD, jobs to be done, user research synthesis.

Create, update, validate, and submit Homebrew Casks. Use when the user mentions Homebrew cask/cask, Homebrew/homebrew-cask, adding a new cask, updating a cask, cask token naming, sha256, url verified:, livecheck, zap/uninstall, or when asked to run brew style/audit for a cask.

Write idiomatic, type-safe TypeScript: errors as values, parse-don't-validate, branded/domain types, deep domain modules, and correct-by-construction APIs. Use when designing or reviewing TypeScript specifically — Result types, tagged errors, branded types, smart constructors, schema parsing, module/import layout, JSDoc, or the cast/`any`/`!` discipline. For language-agnostic design use the architecture skill; for lint/tsconfig config use mechanical-enforcement; for test strategy use testing.

Control UTM virtual machines on macOS programmatically — create and configure VMs via AppleScript, manage lifecycle with utmctl, run commands and transfer files in guests, and set up Windows 11 ARM for headless automation (SSH, not the guest agent). Use whenever the user mentions UTM, utmctl, VMs on Apple Silicon, creating/cloning/scripting a virtual machine on a Mac, or wants to run commands inside a local Windows or Linux VM — even if they don't name UTM explicitly.

A second alpha, colliding with repo/alpha to exercise precedence.

The alpha skill, used for resolve and pointer tests.

Beta skill with a multiline description that must be flattened to a single line for display.